The third and final part of the article series “Building an Azure Analysis Services Model on Top of Azure Blob Storage” has been published on the Analysis Services team blog. It shows that an Azure Analysis Services S9 server can successfully ingest more than 1,000 source files in a 1-terabyte TPC-DS data set stored in Azure Blob Storage in less than 12 hours. Among other things, Part 3 demonstrates how to analyze the memory consumption of the model and how to optimize model size and processing times. This part also discusses the advantages of more sophisticated data sources, such as Azure SQL Data Warehouse, over Azure Blob Storage for reducing the data volume that must be transferred to Azure Analysis Services for processing. Thanks to the modern Get Data experience, you can build a flexible data import pipeline directly in a Tabular model and import even very large data sets with reasonable processing times.

Read the full article “Building an Azure Analysis Services Model on Top of Azure Blob Storage”.